Extended pulmonary resections of advanced thoracic malignancies with support of cardiopulmonary bypass

Surgical resection remains the only potentially curative option for a variety of thoracic malignancies. A radical surgical procedure with cardiopulmonary bypass (CPB) support is rarely offered to patients with advanced thoracic tumors for a variety of reasons. These may include the potential mortality and morbidity, the poor prognosis for subsets of this patient population, or technical considerations. Conventional techniques frequently do not allow for complete resection of advanced pulmonary tumors invading the heart or great vessels. For non-small cell lung cancer invading the mediastinum with infiltration of the great vessels complete resection has been reported in only 30–40% of such cases [1,2]. When the left atrium is affected, partial clamping associated with hemodynamic instability and tumor embolisation [3,4] can be avoided. Complex cardiac resections or reconstructions, replacement of the thoracic aorta, or the common pulmonary artery can only be approached with circulatory support by CPB (Fig. 1). The application of CPB provides complete inspection of infiltrated cardiac or vascular structures, allowing for safe resection margins and intraoperative microscopic control of complete resection. Sarcoma infiltrating the pulmonary arteries: CT-scan of patient 12 presenting with a leiomyosarcoma after induction chemotherapy. Infiltration of the upper, middle and lower lobe of the right lung, pulmonary artery trunk, and right and left main pulmonary artery. The application of CPB for extended pulmonary resections, however, is debated controversially. Following lobectomy or pneumonectomy, respiratory complications occur with an incidence of up to 49% [5]. Prolonged time of anesthesia, longer operations, and cardiac disease have been reported to increase the risk for pulmonary complications following lung resections. In case of additional application of CPB, more frequent and more severe pulmonary injuries have to be expected. Extracorporal circulation is known to cause lung injury and may be harmful, especially with prolonged use. Increased pulmonary complications and excessive bleeding have been observed in concomitant pulmonary resections and cardiac procedures with CPB support [6,7]. The inflammatory response to CPB includes the activation of the complement and coagulation systems, activation of the fibrinolytic and kallikrein cascades, and activation of neutrophils with degranulation, protease enzyme release, production of oxygen radicals, and the syntheses of various cytokines from mononuclear cells (including tumor necrosis factor, interleukin-1, and interleukin-6) [8,9]. Intraoperative pulmonary edema, reperfusion injury, acute lung injury (ALI), and acute respiratory distress syndrome (ARDS) are known to be associated with the use of CPB. The adverse effects of CPB resulted in increased postoperative morbidity, prolonged ventilatory support, and hospital stay [8]. Due to the rare nature of patients undergoing extended pulmonary resections with the application of CPB for mediastinal involvement, only case reports and recently a few small series had been published [10–12]. Prior reports suggested the technical feasibility and in some of the cases, patients were reported to demonstrate an excellent outcome following complete resection. Some authors advocated the use in lung cancer patients; others restricted the indication. In some reports, palliative indications for relieve of symptoms were included. Appropriate indications for extensive surgical treatment in advanced thoracic tumors invading the heart or great vessels have remained not well defined, especially the roles of neoadjuvant or adjuvant radiation or chemotherapy have not been outlined very well. The aim of this study was to analyze indications, techniques, morbidity, mortality, and long-term results of this radical surgical approach. This series adds a further experience to a rapidly expanding field of interest. In a retrospective study, resection of thoracic tumors involving the lung and cardiac structures or the great vessels performed with the support of extracorporal circulation were analyzed. From 8/93 to 8/03, 13 patients underwent an extended pulmonary resection involving cardiac structures requiring cardiopulmonary bypass at the Division of Thoracic and Cardiovascular Surgery, Hannover Medical School, Germany. Patients with cardiac and mediastinal tumors (no pulmonary resection) were excluded. Data on preoperative status, patients’ demographics, preoperative therapies, tumor characteristics, surgical procedures, postoperative complications, and outcome were collected (Table 1). Follow-up data were available from all patients who were included. Patients’ characteristics and pulmonary procedures Patients were discussed in a multidisciplinary conference before undergoing resection. All resections were done in a curative intention. Preoperative evaluation included ruling out metastatic disease and advanced lymph node disease. The indication to offer surgical therapy was made when no alternative treatment was available and based on the individual situation of the patient. Four patients underwent neoadjuvant chemotherapy and radiation prior to surgery. All patients suspected with primary lung carcinomas underwent cervical mediastinoscopy. Diagnostic bronchoscopy was mandatory. If pneumonectomy was planned, a ventilation and perfusion scan was performed. In selected patients, the maximal oxygen uptake, pulmonary artery pressure, and pulmonary vascular resistance were measured. A coronary angiography was performed in all cases. Patients had to present with acceptable cardiopulmonary function allowing for an extended surgical procedure. Diagnostic procedures comprised computed tomography (CT), magnetic resonance imaging (NMR), transthoracic and transesophageal echocardiography. When indicated endosonography of the esophagus, or cine-NMR of the heart, was included. In addition to the preoperatively planned use, CPB was applied in case of unexpected intraoperative findings (Table 1). For statistical analysis the SPSS statistical package (SPSS Inc., Chicago) was used. Cumulative survival was calculated by the method of Kaplan–Meier. Thirteen patients (eight females and five males) were included in the study. Mean age was 47 years (range, 28–64 years). In 12 of 13 patients, the diagnosis of the tumor etiology was established prior to resection. Nine patients presented with a primary tumor of the lung, and three with secondary pulmonary tumors. Three patients had a recurrent laryngeal nerve palsy, and in two cases concomitant phrenic nerve palsy was found. Tumor masses occluded the lumen of one main bronchus in five patients. Prior to surgery, six patients received chemotherapy, in two cases with additional radiotherapy. In eight patients, a tumor of mesenchymal origin was found (Fig. 1). The diagnoses were leiomyosarcoma (n = 3), rhabdomyosarcoma (n = 2), spindle cell sarcoma (n = 2), and a malignant hemangiopericytoma (n = 1). Three sarcomas were originating from the pulmonary artery. In three patients, the histological diagnosis was non-small cell lung cancer (T4). One patient received neoadjuvant chemotherapy and radiation for downstaging in T4, N2 disease. In one case, the tumor was a metastasis of a rectal cancer that had been completely removed 6 years before in a patient without evidence of other metastasis. In one patient, a large mediastinal mass compressing the heart proved to be a large benign bronchiogenic cyst. Four patients presented with incomplete resections (R1) or abrogated prior procedures in institutions without the availability of CPB. The requirement for CPB was preoperatively planned in 9 of 13 (77%) cases. Extent of the tumor (n = 2), or injury of the common pulmonary artery (n = 2), including a case with the benign cyst, required unexpected application of CPB (Table 1). The tumors were approached (Table 1) via a median sternotomy (n = 7), postero-lateral thoracotomy (n = 4), or a bilateral clamp-shell incision (n = 2). For arterial canulation, the ascending aorta (n = 9), the descending aorta (n = 1), or the femoral artery (n = 3) were used. Venous access was bicaval (n = 9), femoral (n = 2), or via pulmonary artery (n = 2) canulation. Mean time of operation was 304.6 ± 112.2 min (153–495 min). Time on bypass (Table 2) was 140 ± 88 min (24–320 min). In five cases, cardiac arrest was induced by cardioplegia with cross-clamping times of 101 ± 37 min (37–145 min); in another five patients, electrical-induced fibrillation was applied. Lowest rectal temperature was 27 ± 3 °C (20–32 °C). An average of 7 (0–18) units of packed red cells, 5 (0–11) units of fresh frozen plasma, and 0.6 (0–2) units of platelets were administered intraoperatively. Cardiac procedures: the majority of tumors affected several cardiac structures Pulmonary extension of tumors required pneumonectomy in nine patients, a right lower bilobectomy in one case, and lobectomies with additional wedge resections from another lobe in three cases (Table 1). In the majority of patients, several mediastinal and cardiac structures were affected simultaneously (Table 2). The left atrium was most often invaded by tumor (n = 9), in two cases extending into the pulmonary veins from the opposite side. The common pulmonary artery (n = 7), the right atrium (n = 3), the right ventricle (n = 2), the aorta (n = 1), and the coronary arteries (n = 1) were resected en-bloc with the pulmonary tumor. For reconstruction of cardiac structures, autologous pericardium was used when available (n = 6). In addition, aortic and pulmonary homografts (n = 4) and polytetrafluoroethylene (PTFE) prosthetic material (n = 4) were used. Special emphasis was taken to verify complete resection by intraoperative frozen sections for immediate pathological evaluation. In one patient presenting with non-small cell lung cancer, mediastinal lymph nodes (pN2) infiltrating the trachea prevented complete resection (Table 2). However, resections proved to be complete resections (R0) by pathological examination in 12 of 13 patients (92%). The 30-day mortality was 15% (2 of 13 patients). One patient died intraoperatively due to right heart failure and pulmonary injury after an excessive time (289 min) on CPB for a complex cardiac reconstruction involving both coronary arteries. Mean time on ventilator in the 12 remaining patients was 30.3 h (3–192 h) (Table 3). Three patients required prolonged times (>24 h) of mechanical ventilation due to acute lung injury. One of these patients died due to development of multi-organ failure on day 4. Two patients recovered from respiratory failure. In two cases, postoperative episodes of pneumonia were treated. One of these patients presented with recurrent atelectases. Other complications observed were two re-thoracotomies for a hematothorax. Mean time on ICU was 2.3 days (1–9 days) and mean time in hospital was 19 days (9–40 days) in 11 patients discharged (Table 3). Outcome and complications Comparing the nine procedures with pneumectomies to the four procedures with lobectomies, the group of pneumonectomies had longer times on bypass (164 ± 94 min vs 87 ± 22 min), longer times of operation (321 ± 132 min vs 270 ± 14 min), and longer periods of postoperative mechanical ventilation (42 ± 61 h vs 13 ± 4 h). More pulmonary complications were observed in pneumectomies (4 of 9 patients vs 1 of 4 patients). The 30-day mortality rate was 22% (2 of 9 patients) following pneumonectomy, whereas no patient undergoing lobectomy was lost. At follow-up examinations no local recurrence of tumor at the site of surgery was detected, following complete resection of the malignancy. Cumulative survival rates for all 13 patients were 62% at 1 year and 53% at 3 and 5 years. After a follow-up of 39 months (0–216 months), seven patients were alive (Table 3). Five of the seven survivors were in class NYHA II and two were not limited in their physical activity. Cause of late death (n = 5) was distant metastatic disease in all cases. All three patients presenting with non-small cell lung cancer died at 4, 6, and 16 months, respectively. Mean survival in these patients (n = 3) was only 8.7 ± 3.7 months, compared to 156 ± 35 months in patients (n = 8) with sarcomas. The 1-, 3- and 5-year survival rate was 62.5% in patients presenting with sarcoma, including two perioperative deaths (Fig. 2). Five of the six patients discharged were currently alive. Two underwent chemotherapy for late development of a single distant metastasis. Survival after extended pulmonary resections with CPB: long-term outcome of patients presenting with sarcoma (n = 8) and non-small cell carcinoma (n = 3), including two perioperative deaths. Kaplan–Meier analysis of cumulative survival. Due to the rare number of patients presenting with thoracic malignancies requiring extended lung resections with CPB support, we identified 13 consecutive cases over a period of 10 years. It had been estimated that less than 0.1% of all thoracic resections were done with CPB [11]. The indication for the surgical radical therapy was based on the individual situation of the patients after interdisciplinary evaluation and discussion of treatment options. Distant metastatic disease or extrathoracic sites of disease had been excluded. All elective procedures were done in curative intention; no palliative indication was considered acceptable in these advanced tumors. In 30.8% of patients, the decision of CPB application was made intraoperatively. In two cases emergent institution of CPB was required for intraoperative injury of the great vessels. Byrne et al. [11] reported 43% of all CPB procedures to be performed for emergency situations. These data underline the importance of CPB availability when advanced thoracic malignancies are surgically approached. It provides a safety net and is helpful when the procedure is more extended than preoperatively expected. In case of suspected tumor invasion of the great vessels of the heart, patients should be directed to a surgical center that is able to provide CP. Four out of 13 patients in our series underwent prior operations in other hospitals that were abrogated due to the unexpected invasion of the heart or great vessels. Park et al. [12] described 3 of 10 patients who underwent CPB-assisted operations following previous unsuccessful attempts of tumor resection. Often patients will benefit from these radical procedures only if complete microscopic resection can be achieved. We found that 92% of patients had an R0 resection in our series. Comparable results (79–86%) have recently been described [10,11]. Considering the high-risk nature of the performed procedures, a perioperative mortality rate of 15% (2 of 13 patients) is not unexpected. This rate is not different to other series reporting 7% (1 of 14 patients) and 11% (2 of 19 patients), respectively [10,11]. To achieve complete resections, we accepted prolonged times on bypass. This underlines the complexity of the procedures. Compared to other studies published, mean time on bypass (140 min) in our experience was longer compared to other reports (87, 99, and 111 min) [10–12]. Time on cardiopulmonary bypass is known to be a crucial factor for patients’ survival in cardiac surgery, when applied for prolonged periods (>3 h). Beyond 90 min of CPB, an increased pulmonary dysfunction has been observed. However, the application of CPB might cause only minimal detrimental effects, if applied for short periods of time, thus allowing for extended pulmonary resections. Surgical interventions for advanced thoracic malignancies are associated with a substantial perioperative risk and frequent postoperative pulmonary complications [6]. For standard pneumonectomies, a mortality rate of 5.4% with a morbidity rate of 59% was reported by Algar et al. [13]. In their series of 242 patients undergoing standard pneumonectomies, pulmonary complications were present in 14% of cases. In other reports, the incidence of pulmonary complications following pneumonectomies ranged from 11 to 49%. Extended pneumonectomies and prolonged anesthetic time were associated with the highest rates of pulmonary complications [5,12]. Acute lung injury and acute respiratory distress syndrome represent severe forms of pulmonary failure, which occur in 2.2–4.2% of all pulmonary resections for bronchiogenic cancer and are associated with a mortality rate of 40–72%. There is considerable evidence that CPB is associated with deterioration of pulmonary function by the of pulmonary edema, pulmonary and pulmonary vascular resistance More frequent and more severe complications with extended periods of mechanical ventilation are if CPB is applied. In our we postoperative pulmonary complications in 5 of 13 patients One of these patients died intraoperatively due to right heart failure and pulmonary after undergoing a complex procedure with prolonged time on CPB. An extended mechanical ventilation period was observed in three patients (>24 h) from acute lung injury. One of these patients and died due to the development of multi-organ failure. All other patients recovered from pulmonary The 30-day mortality rate was 15% (2 of 13 patients). et al. reported on four patients presenting with tumor invasion of only the left who had a short time on CPB ± min), and were for h with no pulmonary complications observed. In a prolonged ventilatory support of days) was described by et al. In their study, the hospital mortality was 11% with complications reported in of patients, including pneumonia and Byrne et al. [11] found a mean of ventilator support of One of the 14 patients an acute lung injury, four patients from pneumonia and three underwent is another frequent following CPB-assisted procedures. In a series of lung resection cardiac operations with CPB, bleeding complications were reported to of patients (4 of 19 patients) In for a following a standard pneumonectomy only in of cases in a by et al. [13]. In our of 13 patients underwent for perioperative with fresh frozen and Other series reported an incidence of postoperative bleeding complications following lung resections with CPB of and respectively [10–12]. Extended pneumonectomy with CPB is a rarely performed procedure. Nine patients of our series underwent extended pneumonectomy with CPB for tumor invasion of cardiac structures, compared to four patients, who had a et al. included four patients with sarcomas in their study not this to patients with less extensive pulmonary resections. et al. reported on Byrne et al. [11] on and Park et al. [12] on extended In addition, several case reports had been published in the In addition to the in the of pulmonary the group of patients with pneumonectomies had longer times of operation and longer times on bypass. These intraoperative were associated with extended time periods of mechanical ventilation and with rates of morbidity and patients undergoing pneumectomies pulmonary mortality rates and more frequent pulmonary complications have been reported with pneumonectomy compared to other less extensive pulmonary resections Acute lung injury has been reported to occur often in pneumonectomies in lobectomies in standard resections However, has been no published on increased complications in extended pneumonectomies with From our we that extended lobectomy with CPB is a safe pneumonectomy in this remains to be associated with sternotomy is the most often used to thoracic and cardiac procedures. For some of the complex cases, we found a clamp-shell used in pulmonary to offer the should be In case of tumor femoral or pulmonary artery can be alternative of the extracorporal bypass. of the tumor and the structures should be prior to of CPB to the time on CPB. However, if tumor or is cardiac arrest or fibrillation should be prior to tumor We with is known to postoperative and cause pulmonary bypass to activation of and of the activation of these inflammatory with a of pulmonary dysfunction of and the inflammatory of CPB and the ventilatory following the cardiac surgery CPB units and with and less are known to the inflammatory effects and cause less pulmonary The inflammatory response can be when from the field is not recovered into the CPB a is used. In addition, with protease or complement might further the activation of and other in the inflammatory of CPB. For the can be that these technical will the safety of this procedure and the risk of pulmonary in patients with advanced non-small cell lung carcinomas invading the mediastinum is Survival in carcinomas that were resected without CPB has been reported to be after 3 years and after 5 years in another study A hospital mortality rate of and a 5-year survival of in hospital survivors were reported by et al. Park et al. [12] that is to complex thoracic resections with the potential of prolonged and substantial morbidity and mortality long-term survival is However, some patients especially with disease were reported to have survival rates following complete resection The indication for surgical treatment in non-small cell lung cancer should be radical surgical therapy may be considered if patients present with disease. The benefit of a neoadjuvant chemotherapy and radiation with following induction which is for patients with advanced non-small cell lung cancer, remains to be in this procedures with CPB in palliative reported do not to be [10,11]. In to non-small cell pulmonary sarcomas have a different of disease and indications for surgical therapy are Following the resection of primary pulmonary 5-year survival rates of have been reported In our primary and secondary pulmonary sarcomas invading the and the heart were resected with CPB. The 1-, and 5-year survival rates were respectively (Fig. 2). These results underline the of complete surgical resection a primary in the treatment of pulmonary sarcomas. advanced sarcomas may have a long-term chemotherapy should be considered in advanced sarcomas. We found that neoadjuvant chemotherapy in two patients that were at for partial response and complete surgical resection. et al. reported on three patients undergoing preoperative chemotherapy are not data to support the use of induction therapy However, adjuvant therapy is indicated in these rare malignancies. The rate of recurrence of pulmonary sarcomas at the primary site or distant metastasis following resection has been reported to be The experience described is limited due to a small number of patients, not allowing for statistical evaluation. However, our data support the feasibility of surgical therapy with CPB for pulmonary malignancies invading cardiac We to radical resections with support of CPB. patients benefit from complete resections with long-term survival especially patients presenting with advanced It may be that if pulmonary resection is limited to lobectomy and the time of CPB is no increase in pulmonary complications will have to be expected. However, extended pneumonectomies with CPB support to be associated with a substantial risk of pulmonary